Skip to content

Dead Hard Drives 2

I should have thought of this…

I did a quick search on the Internet for “hard drive platter transplant” and it lead me to a link on hackaday about doing such an operation. The article itself seems to be crap; the important part are all the comments left. There are a few people who seem to know what they were doing. (The cryptic “freezing the hard drive” makes no sense and it seems most people talking about have no idea what it does.)

Though it is unclear at this point exactly why my platter transplant failed, here are some ideas:

1. During normal operation, the platters are fixed in position relative to each other. Thus any data written to the drive most likely depends on this relative position. In this particular hard drive, the only way to remove the platters without altering this alignment would have been to remove them with the motor, which is screwed to the chassis of the drive via three screws that sit directly beneath the third platter—in other words, milling the chassis on the back side until you get to the motor flange would have in theory released the whole assembly together, at which point maybe (in theory) you could have dropped the three screws out, tapped those holes with a thread, then drill the appropriate clearance on the second chassis (with the working head) so you could screw the motor in from the outside. This is a very sketchy plan, but it may actually work.

2. If indeed there is a “calibration” between the head and the platters, then by loosening any component inside the drive you are screwed. Frankly I still find this hard to believe—for example, my car has electronic throttle control so the traction control can close the throttle opening on its own regardless of the gas pedal’s position. I have read that the minute you start the car, the throttle valve opens fully and then closes to idle so that every time you start the angular position sensor is calibrated. It would seem that the hard drive would have some sort of fail-safe system like this—something that would tell the heads, “okay, you’re in parked position now, so scan the platter and look for pattern XXX to know where the hell you are.” Then again, the ^&#^$ should have put some surge protection so you don’t fry the ^%#^$ heads.

3. According to the comments on the hackaday article, and my own intuition, the clean room is not a necessity. Especially after seeing how much wind the platters generate when they are spinning, I can’t imagine that as long as you don’t leave it sitting in a barn for a week how it wouldn’t immediately clean itself upon spin up. (Of course at some point I was blowing a platter with canned air and it suddenly condensated which left a nasty stain on it—probably not good.)

In conclusion—I’m willing to bet it’s the platter alignment issue. I doubt as it was suggested by a friend of mine that you can somehow get some software that reads entire platters blindly rather than trying to read the actual data and then run it through some crazy #$%# to “reconstruct” the data from these images.

In any case any chance of trying it is done as I purposely scratched the hell out of the first platter.

Post a Comment

Your email is never published nor shared. Required fields are marked *